YOLO-Based Deep Learning Model Shows Feasibility for Detecting Lumbar Disc Herniation on MRI

A YOLO-based AI model detected lumbar disc herniation with an overall precision of 0.74 and recall of 0.70 in a small single-center retrospective study. The model remains hypothesis-generating and is not ready for clinical use due to the tiny test set and lack of external valida…

• Design: Retrospective, single-center feasibility study using 372 sagittal T2-weighted MRI slices; test set was only 10 images.
• Key secondary finding: The model demonstrated high recall for L5-S1 herniated discs (0.96) but performance was lower for non-herniated classes, and bootstrap confidence intervals were wide.
• Major limitation: Very small test set (10 images), no external validation, single-center data, and no comparison with radiologist performance.

RadPigeon summaries are original and for information only. They are not clinical advice.